TRANSACTION COSTS AS MEASUREMENT OF INSTITUTIONALLY
LED STRUCTURAL CHANGES
MARIA IVANOVA
Faculty of Economics and Business Administration,
Sofia University “St. Kliment Ohridski”, Bulgaria
This paper aims to bind economic ideas outside of the mainstream with the established
techniques of input-output analysis. New Institutional Economics brings out the concept of
transaction costs as those costs that economic agents pay in order to exchange the results from their
production and thus reap the benefits from the labour division. The process of economic exchange can
be easier or more difficult, that is to say costly, depending on the rules for this activity. “Rules of the
game” in a society are built in institutions. Therefore institutions and institutional change are as
important as technologies and technological change. The paper presents a partial model for analyzing
the results from the interactions between institutional and technological changes that happen at
economy-wide level. It is is a modification of the original framework of Wallis and North (1994)
because it uses nominal data. The input data come from an international database and this allows
reliable inter-country comparisons. The model is tested with 20 countries for two 5-years periods.
Examples of successful and unsuccessful structural changes are given. The model can identify if the
consequences from these different impulses are synchronized and if this led to improvement in the
overall efficiency of the economy. Also, it discerns economies with sluggish sectors that might fail the
whole system in its future development.
Keywords: transaction costs, institutional change, technological change, efficiency
1. INTRODUCTION
Neoclassical microeconomics examines the problems of efficient
transformation of inputs (raw materials, capital, and labour) into new products. The
main postulate is that due to comparative advantages it is more productive if people
and their economic organizations specialize in different businesses and afterwards
exchange the fruits of their labour. The relevant phenomenon here is technological
change and the consequent rise in the productivity of this transformational process.
New institutional economics brings out the concept of transaction costs as
those costs that economic agents pay in order to exchange the results from their
production and thus reap the benefits from the labour division. The exchange
process can be easier or more difficult, that is to say costly, depending on the rules
for this activity. “Rules of the game” in a society are built in institutions. Therefore
institutions and institutional change are as important as technologies and
technological change. In successful societies institutional and technological changes
lead to those structural changes in the economy that lower all kind of costs,
transformational and transactional (Wallis and North 1994).
In 1986 Wallis and North, in their article “Measuring the transaction sector in
the American economy, 1870-1970”, try to give shape to the concept of transaction
costs estimating their size as a proportion of the total economic activity in an
economy. They gain several followers1 who replicate the measurement methodology
using data for different countries, but little is achieved in the field of testing
hypotheses or empirical approbation of models of the new institutional economics on
macroeconomic level. All efforts remain in the field of the descriptive comparison.
The objectives of this paper are three. The first objective is to adapt a part of a
framework for integration of institutional and technological changes (Wallis and
North, 1994) so it can be used with easily accessible and standardized nominal data.
The second objective is to process international data according to this adapted model
and see if they reveal different patterns. The third objective is to define and find
examples of successful structural changes in the economy using these data and this
model.
The remainder of the paper is structured as follows. Section 2 explains the
concept of transaction costs, their connection with institutions, and presents shortly a
model of interplay between institutional and technological changes in the economy,
as North and Wallis (1994) propose it. Section 3 gives reasons for the used data set,
and describes it. Section 4 links this particular data set with part of the
aforementioned but now adapted model. Section 5 reports the calculation
methodology, and Section 6 discusses the results. Section 7 concludes and
summarizes.
2. TRANSACTION COSTS, INSTITUTIONS, INTERACTIONS
2.1. Transaction costs
On the macroeconomic level transaction costs are vaguely defined. This does
not mean they are unimportant. When something sounds vague, it is hard to be
measured. When it is not measured, it is not possible to test it and the theories and
the models connected with it. When a theory is not clashed with the empirics, it
cannot develop, and it cannot prove as an important theory and enter the
mainstream of economics. In this section I give a simple example of what transaction
costs are, how they are influenced by institutions, and why institutional changes are
worth keeping an eye on them.
In economy-wide aspect transaction costs are cited as “costs of running the
economic system” (Arrow, 1969) and then every researcher plunge into their own
1For
Australia – Dollery and Leong (1998), for Argentina – Dagnino-Pastore and Farina (1999), for
New Zeeland – Hazledine (2001), for Poland – Sulejewicz and Graca (2005), for Bulgaria - Chobanov
and Egbert (2007), Egbert, et al. (2010).
comprehension of the matter. I prefer to take my stand from the words of Steven
Cheung for whom transaction costs are “all those costs that cannot be conceived to
exist in a Robinson Crusoe economy where neither property rights, no transactions,
nor any kind of economic organization can be founded” (Wang, 2007). In
microeconomics textbooks we can find the Robinson Crusoe economy as an example
of how the production possibilities frontier can expand after Crusoe starts to trade
with Friday. If Crusoe is better at doing one thing, say picking nuts, and Friday is
better at doing another thing, say fishing, then for both of them it is a better solution
to specialize in one activity and exchange the surpluses. And that is how the benefits
from division of labour due to comparative and absolute advantages are introduced
to students.
Now imagine that Crusoe and Friday do not trust each other while making
business together. A situation that is typical for big markets with infinite possible
contractors with never repeating contacts. Being only two on a desolate island almost
precludes any idea for cheating and opportunistic behaviour but imagine that a ship
of H. M. the Queen of Britain has discovered Crusoe and Friday and the two
islanders have started trade with the British Empire. This is the only way to increase
the consumption of the nuts and fish but it also imposes unpleasant consequences
like a representative of the Empire being sent on the island. He, or nowadays she,
will play the role of a law maker, a judge, a policeman, a soldier, a fireman, a builder
of public infrastructure, a public clerk and a producer of all other public services that
a developed economy needs in order to regulate the infinite number of transactions
between agents that may not have good reasons to behave honestly, or perfectly
rationally.
2.2. Institutions
The presence of this judge-policeman-clerk can actually be “pleasant” or
“unpleasant”, that is to say in economic terms “efficient” or “inefficient”. If his
presence increases the production of nuts and fish well enough to satisfy at higher
level the needs of now three people in the insular economy, then the institutions that
rule this economy are efficient. In the New Institutional Economics institutions
encompass a wide range of rules that guide the economic behaviour of the agents.
Starting from inner believes and values to legally set norms. If agents tend to cheat,
or if laws are made equivocal, the Crusoe-Friday economy will need more
representatives from the Empire to punish the cheaters and to explain the laws, and
this will have no direct effect on the ability of Crusoe to pick nuts and the ability of
Friday to catch fish. There will be no technological change in the economy, no change
in the resources needed for production. There will be more transaction costs – the
reward for the judge-policeman-clerks – because of this nature of the institutions.
More costs to perform the economic transactions and same costs to transform the
resources (fish in the sea) into products (fish in the pan), this is not a successful
development of the economy. More transaction costs, combined with bigger drop in
the transformation costs, would be another story.
2.3. Interactions
In their paper “Integrating Institutional Change and Technical Change in
Economic History: A Transaction Cost Approach” (1994) Wallis and North develop a
framework for analysing the relationship between technical and institutional
changes. Technical change, which I call here technological, is any change in the
composition of inputs that are transformed to products for final use or for
intermediate use in further stages of the production process. In my example Crusoe
and Friday perform such transformation function in the economy and all resources
used for it are transformation costs. Their labour, the capital they may use, the raw
inputs like bait for the fish. Institutions set the level of transaction costs – labour,
capital, raw materials for the work of the judge-policeman-clerk, who has a
transaction function in the economy. Institutional change may increase or decrease
the transaction costs but more important is how these changes interact with the
changes from the technological innovations. Altogether they may lead to increase in
productivity, and then bigger costs will be compensated with even bigger output.
Second point in the paper is that both types of changes can give impulse in the
economy. For example the legal construct of the corporate business, once invented,
allowed for the collection of huge financial resources from multitude of capital
owners with limited responsibility who do not need to know each other or to run
directly the company. The financial securing of a production on large scale made
reasonable the invention and implementation of technologies for mass production.
Of course, it is vain to think we can undoubtfully separate the two processes and put
one of them in the beginning: is that first engineers invented the new technologies or
that first institutional phenomena took place in the society. It is good enough if the
two types of changes happen synchronously with a reasonably small lag in time so
that potential benefits from new production technology are not suffocated by
inappropriate exchange mechanisms. It is more about the flexibility of the
institutions and the rates with which transaction costs change, not their absolute
levels. In fact, we cannot judge the efficiency of transaction costs when institutions
are at rest because we cannot measure an alternative “what if” state. We can only
assess the efficiency of the direction in which they change. Do they lead to bigger
productivity or not.
Third point in the paper is that because innovations in one industry spill over
the economy through the intermediate consumption, thus technological and
institutional changes interact. For example when Alexander Bell invented the
telephone, this was a technological change in its own industry. The industry of
telecommunications has a transformation function in the economy. It combines raw
inputs in such a way that we can transmit information over long distances in order to
communicate. The invention of the telephone is called by Wallis and North “a
transformation augmenting technical change”. It augmented the transformation
function of the telecommunication industry. With the same resources, combined in a
new way, we can transmit more quantity of information, and at a better quality. For
the other industries, which use the telephone to coordinate the complex net of the
modern production process, the invention of the telephone is a transaction
augmenting technical change. It supports the increase of productivity of the
transaction function in the companies. Now, we have to distinguish types of
transaction costs.
In brief, Wallis and North (1986) discern three parts of the transaction function
in the economy. There is one that is performed by specialized intermediaries, like
financial industry, insurance industry, real estate industry, wholesalers and retailers.
The companies in these industries help other industries sell their products. The
essential feature of the value added of these industries is the change in the property
rights, the change in the legal status of the exchanged products. Second place to find
such a function is within the companies from the transformation industries. There is
a multitude of professions which single purpose is to sustain the transformation of
inputs from the entrance of the firm to its exit. If Robinson Crusoe was to set up the
company Crusoe Inc. in order to reach economies of scale, he would have to deal
with secretaries, foremen, managers, clerks, accountants, marketing specialists, etc. –
all of them people with no direct link to the process of picking nuts from the palms.
We cannot say outright that they are all useless. Third, here comes the Government
like a big traffic policeman. It is supposed to relieve all tensions that occur in the
process of exchange of products where specialized intermediaries (because of market
failure) or specialized professions (because of principal-agent conflict) cannot cope
with.
The ambition in this paper is only the first type of transaction function to be
modelled with input-output tables. It is the type that concerns spillovers from
institutional or technological changes through the economy using the linkage of
intermediary consumption. Wallis and North (1994) imply eight types of changes,
though they do not give historical examples for all of them. First pair, by type of the
impulse, they are a technical or an institutional change. If it occurs in the industries
with transformation function, it is a technological change; otherwise it is an
institutional change. For example the development of investment banking in the
financial sector is an institutional change. Second pair of possibilities, is by type of
the influenced economic function. If the change influences the transformation
function in the economy – this is the ability to transform inputs into outputs – this is
a transformation changing innovation. Otherwise it is transaction changing
innovation. The third pair is whether the change augments or attenuates the results
from the economic performance. The eight possible types of changes, following from
the interactions between institutional and technological development, result from
these 2 x 2 x 2 permutations.
Last but not the least to mark off, Wallis and North (1994) speak of
productivity which presumes data in constant prices while more practical from
empirical standpoint it would be to formulate a model in nominal terms.
3. DATA SET
It is one thing to model in constant prices, and another to use what life can
give us. The concept of productivity requires use of data at constant prices. But
input-output tables at constant prices are not often met, and especially for free in
internet. Nominal data would enormously relieve the comparisons in time and space.
The original paper from which the entire endeavour to quantify the
transaction costs at macro level started is Wallis and North’s “Measuring the
transaction sector in the American economy, 1870-1970”. It uses USA specific
statistical data sets and ever after that, when replicated, there are doubts that
differences in the results might be explained with differences in the statistical
methodologies for creating the data sets. On the other hand, the institutionalist
theory for transaction costs needs testing in different contexts and international
comparisons are desired. That is why I look for an international database with
uniformly created data.
Third, institutions evolve, on general, slowly. A timespan of a century is the
most comfortable background for institutional analysis, like in Wallis and North
(1986). But this means, on the current stage of history, to compare only developed
economies with long statistical records. This disadvantage of less developed or
transitional economies may be offset by the fact that, stimulated by globalization,
they evolve more rapidly than established market economies. Therefore institutional
and technological changes may show up in changing transaction and transformation
costs in shorter periods.
The data for this research is from the Organization for Economic Co-operation
and Development. STAN benchmark year, industry-by-industry, input-output total
tables for 20 counties from mid-1990s to mid-2000s are used. I could have used tables
for more countries if I have limited the search only to two points in time but I want to
have three dates so that to be able to check the path of development from one 5-years
period to another 5-years period. I also exclude countries that have different number
of industries in different points in time. That is why I use only the tables for those 20
counties which have information for mid-1990s, early 2000s and mid-2000s and keep
the number and types of industries unchanged.
4. MODEL FOR INTERACTIONS
TECHNOLOGICAL CHANGES
BETWEEN
INSTITUTIONAL
AND
How can we adapt the example of Alexander Bell’s telephone into a model
that uses nominal data?
First, I introduce several notations borrowed from Miller and Blair’s textbook
(2009).
Zij:
monetary values of the transactions between pairs of industries,
from each industry i to each industry j
Xj:
monetary value of output of each industry j
Aij = Zij/Xj: technical coefficient, or direct input coefficient
TF ind:
industry with mainly transformation function (TF) in the
economy
TA ind:
industry with mainly transaction function (TA) in the economy
BL:
backward linkages
Here I present two possible scenarios for the economy-wide results from the
invention and the introduction of the telephone in the production process. The first
case is presented in Table 1. The telecommunication industry is the first TF industry
in the Z table. We recognize the presence of an adopted innovation by the fact that it
is expensive and demanded and that is why its Zij increases in relation to previous
period. The output of the telecommunication industry increases too. The telephone is
adopted in the production process but first, this does not lead to equal increase in the
output of other industries, and second, the transformation and transaction functions
do not change. This means that to serve 10 % increase in the output of any industry,
10 % increase in the Zij’s is needed both from supplying TF industries, other than the
telecommunication, and from supplying TA industries. The sum of the technical
coefficients “from TF ind to TF ind” and “from TF ind to TA ind” will, in general,
rise. The sum of the technical coefficients “from TA ind to TF ind” and “from TA to
TA ind” will stay the same. The backward linkages in the economy will rise. In the
terms of nominal data, this is what we can define as an unsuccessful change. The
economy has become more inefficient. It may produce more output but for it the
economy uses even more intermediate resources.
Table 2 presents the second case when because of the technical change
Alexander Bell’s telephone there will be augmentation in the transaction function of
the economy. All those intermediaries – specialized in controlling the cost of
purchasing inputs, monitoring the production process, selling outputs, and securing
the exchange of property rights in the presence of opportunistic behaviour and
bounded rationality – will be able to supply their output with greater efficiency.
Table 1. Links between transaction matrix (Z table) and technical coefficients
matrix (A table): technical innovation with no effect on economic functions
Z-table change
TF ind
1.5
1.1
1.1
1.1
1.1
1.1
1.1
TF ind
1.5
1.1
1.1
1.1
1.1
1.1
1.1
TF ind
1.5
1.1
1.1
1.1
1.1
1.1
1.1
TF ind
1.5
1.5
1.5
1.5
1.5
1.5
1.5
TA ind
1.5
1.1
1.1
1.1
1.1
1.1
1.1
TA ind
1.5
1.1
1.1
1.1
1.1
1.1
1.1
TF ind
TF ind
1.36
TF ind
1.36
TF ind
1.36
TF ind
1.00
TA ind
1.36
TA ind
1.36
TF ind
TF ind
TF ind
TA ind
TA ind
BL
1.00
1.00
1.00
1.00
1.00
1.06
1.00
1.00
1.00
1.00
1.00
1.06
1.00
1.00
1.00
1.00
1.00
1.06
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.06
1.00
1.00
1.00
1.00
1.00
1.06
TF ind
TF ind
TF ind
TF ind
TA ind
TA ind
X
A-table change
Table 2. Links between transaction matrix (Z table) and technical coefficients
matrix (A table): technical innovation with positive effect on transaction function
Z-table change
TF ind
TF ind
TF ind
TF ind
TA ind
TA ind
X
TF ind
1.5
1.1
1.1
1.1
1
1
1.1
TF ind
1.5
1.1
1.1
1.1
1
1
1.1
TF ind
1.5
1.1
1.1
1.1
1
1
1.1
TF ind
1.5
1.5
1.5
1.5
1.4
1
1.5
TA ind
1.5
1.1
1.1
1.1
1
1
1.1
TA ind
1.5
1.1
1.1
1.1
1
1
1.1
A-table change
TF ind
TF ind
TF ind
1.36
1.00
TF ind
1.36
1.00
TF ind
1.36
1.00
TF ind
1.00
1.00
TA ind
1.36
1.00
TA ind
1.36
1.00
TF ind
TF ind
TA ind
TA ind
BL
1.00
1.00
0.91
0.91
1.03
1.00
1.00
0.91
0.91
1.03
1.00
1.00
0.91
0.91
1.03
1.00
1.00
0.91
0.67
0.93
1.00
1.00
0.91
0.91
1.03
1.00
1.00
0.91
0.91
1.03
The output of the industries may rise but this will not require equal increase in the
inputs from the TA industries. The sum of the technical coefficients “from TF ind to
TF ind” and “from TF ind to TA ind” will, in general, rise. The sum of the technical
coefficients “from TA ind to TF ind” and “from TA to TA ind” will decrease. The
backward linkages in the economy may stay the same, or move up or down. It
depends on the dominance of the effect from Atf,j’s or from Ata,j’s.
Several notes have to be made here.
In Table 1 we may not deal with an Innovator industry. It may be a
Monopolist industry. Of course, every innovator is in the beginning a monopolist,
before it’s innovation is copied by competitors. But when working with transaction
costs and institutional change, new institutional economists paint on a larger canvas.
The technological changes are fundamental and they concern the whole industry, not
particular company in it. In this sense, there are no alternatives to bring down their
cost. If they bring positive change for the economy, they will allow other industries
to increase their output with fewer resources. This will drag down the backward
linkages in the economy. If the Innovator industry acts like a monopolist for a long
time, if it just redistributes the value added in the economy to itself, if the transaction
and transformation functions in the economy do not augment, then the backward
linkages in the economy will, in general, increase, showing worsening in the overall
efficiency. In this case I would call the monopolist “a racketeer”.
A racketeering transaction industry may emerge from rules and laws that do
not serve the productive, value-adding, processes but the processes of redistribution.
Take for example the countries from Eastern Europe and the Balkans which left the
communist economic system in the end of 1980’s. They all had to adopt new
institutions to serve the market economy. But some of them are in NATO and EU
now, while others are still outside these unions. And even between EU-members
there are differences in the economic performance though they all have passed
through more or less same accession period. EU capital markets are opened which
means that technological changes can spill over the member countries equally2.
Formal rules are on the surface adopted by all member countries but their
enforcement may drastically differ. If we have a quantitative model which catches
the differences in the consequences from institutional change, we may be able to
discriminate good practices from bad practices.
I try to model not innovations in particular industry but as a whole in the
sector of transaction industries versus the sector of transformation industries. This is
done in Table 3 and in this case the issues of different industry shares, magnitude of
particular backward multipliers, level of aggregation (whether it is four TF industries
versus two TA industries or other combinations) will not be severely important for
the working of the model.
Let’s look at the model. We recognize the happening innovations, or call them
just changes, by the increase of the output of the TA sector or the TF sector. Because
the data are nominal and because generally in the economy there is inflation, the
outputs will always increase. But when an industry, or a sector, has the advantage of
the innovator (or the racketeer), the nominal value of its output will outjump the
inflation in the output of the other sector.
If the two sectors can synchronize their development, we cannot distinguish
the first block of results from the third. Just like in real economy, where we can
hardly separate chronologically engineering from social innovations, but ultimately
we are interested in the final result up to some moment in time, here the model does
not discern between types of impulses as long as changes have synchronized during
the examined period.
More importance is given to the influence on transaction and transformation
functions in the economy, how they combine and the eventual effect on the
backward linkages, that is on the efficiency in the economy. Also, such matters like
import over export are indirectly concerned. For example if the TF industries’ output
does not increase while at the same time the TF function attenuates, then the missing
resources will have to be imported in the economy. The same effect, and also
inflation, can follow an increase in the value added of TA industries, which is equal
to increase in the incomes of the people who work there or invest there, if this
increase is not soon enough spilled over to TF industries which provide the goods
and services of ultimate interest. Remember that the business of the judgepoliceman-clerk in the Crusoe-Friday economy is not end in itself. Nuts and fish are
end in itself, and the money in the economy is ultimately chasing them.
Not being an expert on technological spillovers, I feel their success depends on the quality of the
human capital in the countries. So differences in human capital might be one explanatory factor here.
2
Table 3. Interactions between transaction industries sector (TAS) and transformation industries sector (TFS)
the importance of TAS grows
changes in X of TAS and TFS are equal
leads to equal growth in TFS
Atf,tf
Ata,tf
Atf,ta
Ata,ta
BL
no change
no change
no change
no change
down
up
no change
no change
no change
no change
down
up
no change
down
up
down
no change
down
no change
down
down
down
down
up
down
down
down
up
down
not clear
up
no change
up
no change
up
augmentation in TA function
up
down
up
down
not clear
attenuation in TA function
up
up
up
up
up
no change in TF function
no change in TA function
augmentation in TA function
attenuation in TA function
augmentation in TF function
no change in TA function
augmentation in TA function
attenuation in TA function
attenuation in TF function
no change in TA function
does not lead to growth in TFS
(at all or it is smaller)
no change in TF function
no change in TA function
augmentation in TA function
attenuation in TA function
augmentation in TF function
no change in TA function
augmentation in TA function
attenuation in TA function
attenuation in TF function
no change in TA function
augmentation in TA function
attenuation in TA function
1995-2000
2000-2005
Germany
Spain
Belgium,
Estonia
BEST
Estonia
Hungary
change in X of TAS > change in X of TFS
Atf,tf
Ata,tf
Atf,ta
Ata,ta
BL
1995-2000
no change
no change
no change
down
no change
no change
no change
down
no change
down
no change
up
no change
up
up
down
down
down
no change
down
up
down
down
down
no change
down
up
down
down
not clear
Slovenia
up
no change
up
no change
up
France
up
down
up
down
not clear
up
up
up
up
up
2000-2005
France
Slovenia
Netherlands
the importance of TFS grows
changes in X of TAS and TFS are equal
leads to equal growth in TAS
no change in TA function
no change in TF function
augmentation in TF function
attenuation in TF function
augmentation in TA function
no change in TF function
augmentation in TF function
attenuation in TF function
attenuation in TA function
no change in TF function
augmentation in TF function
attenuation in TF function
Atf,tf
Ata,tf
Atf,ta
Ata,ta
BL
no change
down
up
no change
no change
no change
no change
down
up
no change
no change
no change
no change
down
up
no change
down
down
down
no change
down
down
down
down
down
up
down
up
down
not clear
no change
down
up
up
up
up
no change
down
up
up
up
up
up
not clear
up
does not lead to growth in TAS
(at all or it is smaller)
no change in TA function
no change in TF function
augmentation in TF function
attenuation in TF function
augmentation in TA function
no change in TF function
augmentation in TF function
attenuation in TF function
attenuation in TA function
no change in TF function
augmentation in TF function
attenuation in TF function
1995-2000
2000-2005
1995-2000
2000-2005
change in X of TFS > change in X of TAS
Atf,tf
Ata,tf
Atf,ta
Ata,ta
BL
no change
down
up
no change
no change
no change
no change
down
up
no change
no change
no change
no change
down
up
no change
down
up
down
down
down
no change
down
up
down
down
down
down
down
not clear
no change
up
no change
up
up
down
up
down
up
not clear
up
up
up
up
up
5. CALCULATION METHODOLOGY
The industry-by-industry input-output tables use a standard industry list
based on ISIC Revision 3. For all countries except Ireland there are 36 industries in all
three periods. Ireland does not have the industry “C23 Coke, refined petroleum
products and nuclear fuel”.
The transaction industries are:
C50T52 Wholesale and retail trade; repairs
C65T67 Finance and insurance
C70 Real estate activities
C74 Other Business Activities
C75 Public admin. and defence; compulsory social security
This choice is in compliance with the framework presented in Section 2 and with the
framework of Wallis and North (1986, 1994). The level of aggregation introduces
some problems that are fully discussed in Egbert and Ivanova (2009) and in Egbert et
al. (2010). For example the presence of repair services in industry C50T52. The
transaction industry of the trade network is slightly “polluted” by the transformation
function of the repair services. I prefer not to pay attention and to look at the “big
picture” for the sake of international comparisons and easily available data.
All other industries are treated as transformation industries. The essence of
their business is to transform inputs into new products and services in the same way
they would have done it in a Robinson Crusoe economy or in a Crusoe-Friday
economy without the judge-policeman-clerk. It is true that inside these
transformation industries some of their resources perform transaction functions.
These are the secretaries, foremen, managers, clerks, accountants, marketing
specialists from Sub-section 2.3. Transformation industries do not buy all they need
for the exchange of their production from the market. There can be substitution
between transaction services bought from outside (visible in the technical
coefficients) and the transaction services produced inside the firm (partly visible in
the types of occupations of the labour factor). I think such a substitution cannot be
that great in all of the 20 countries so that to stultify a model for interactions between
institutional and technological changes which based only on input-output tables.
Transaction costs are the sum of Zij from the transaction industries.
Transformation costs are the sum of Zij from the transformation industries.
Backward linkages are the sum of transaction and transformation costs.
Change in costs is calculated as the difference between Aij from new period
minus Aij from old period. Because Aij’s are between 0 and 1, their change should be
read as “percentage points change” after multiplying by 100. Change in industry
output is calculated as index number Xj-new/Xj-old.
6. RESULTS AND INTERPRETATION
The results for the first period, from mid-1990s to early 2000s, are given in
Appendix 1. Appendix 2 is for the results from the second period, early 2000s to mid2000s. All graphs present median changes. In this way outliers – single industries
that lie too far away from the central tendency for the economy – are excluded. First,
the changes in the transformation costs by types of industries are given. They are
expressed as changes in the technical coefficients, or changes in Atf,ta and Atf,tf (see
Table 3 for notation). On the second graph, the changes in Ata,tf and in Ata,ta are
presented, or the changes in the transaction costs by types of industries expressed in
technical coefficients. Last, the change in backward linkages for the whole economies
is given. In other words, for example a fall in the transaction costs means that the
technical coefficients Ata,j fall. This is equal to augmentation in the transaction
function in the economy.
6.1. Countries that fit the model
The model recognizes interactions between the transformation sector and the
transaction sector if the changes in one type of costs move in the same direction for
both types of industries, or if there are no changes for both types of industries. On
this ground, in the first graph in Appendix 1 there are eight countries that are eligible
for the model as for the changes in the transformation cost. These countries are
Denmark, Estonia, Finland, France, Germany, Hungary, the Netherlands, and
Slovenia. In other cases, like in UK, the movements are in different directions, and
this means that there is not an economy-wide change in the transformation function.
The transaction sector and the transformation sector develop into their own paths
without interaction between each other.
From the second graph eligible countries are Austria, Estonia, France,
Hungary, Japan, Slovenia, UK. Putting together both types of changes – in
transformation costs and in transaction costs, – the model indicates four countries
with different patterns of interactions between transformation and transaction
sectors. They are marked in Table 3.
Estonia and Hungary have sectors that can synchronize their developments.
The difference in the change of outputs of the two sectors is below 10 % points, which
I take as a mark line. From mid-1990s to early 2000s Estonia experiences
augmentation in the transformation function and attenuation in the transaction
function. For Hungary the augmentation is in the transaction function, in trade off for
attenuation in the transformation function. The model cannot confirm that there are
institutional and technological changes, and what order they happen. It can only
show that if there were such changes, the structure of the economy allows the two
sectors to grow equally and its functioning is flexible – it can trade off attenuation in
one function for augmentation in the other. Still this does not help Hungary from
being the country with the biggest growth in the backward linkages among all
examined countries.
In Slovenia and France the growth in the output of the transaction sector
exceeds the growth in the output of the transformation sector by more than 10 %
points. In Slovenia this is combined with augmentation of the transformation
function and no change in the transaction function. The importance of the transaction
sector grows relatively to that of the transformation sector. This can be interpreted as
growing number of economic transactions that have to be served – larger markets,
more customers, or more regulations. But at the same time industries raised their
productivity as for the transformation process. Maybe it is because of bigger
economies of scale after the enlargement of the market. In France the growing
importance of the transaction sector does not help the industries to increase their
efficiency. Slovenia ends as the only country that has a fall in the backward linkages
while France is among the countries with moderate opposite direction.
We can trace Estonia, Slovenia and France in the next period, from early 2000s
to mid-2000s. The transaction function on economy-wide level stays the same but the
transformation function augments. Maybe this augmentation is due to a
technological innovation and the good news here is that the transaction industries do
not respond in a limiting way. These relationships are of main interest for
institutional economists. As Wallis and North (1994 p. 610) say:
“The framework developed in this paper suggests a new historical perspective on the
relationship between technical and institutional change. It questions whether the growth
enhancing effects of technical change – driving down transformation costs – are ultimately
limited by the rising transaction costs associated with the institutional changes necessary to
implement new technologies.”
The behavior of the Slovenian economy is the same, only that the output of the
transaction sector keeps on growing more than its transformation counterpart. This
may be explained with more exports of transaction services to the rest of the world.
France, after having attenuated its transformation function, now attenuates its
transaction function. Certainly the French path of development is not as bright as
that of Estonia and Slovenia. It actually is not bright at all.
6.2. Countries that “do not fit” the model
What can be said for the countries that do not fall into the categories of the
model?
For example in the second period the Slovak Republic experiences the biggest
fall in the backward linkages. With the original data being in nominal terms, this
means that the Slovak entities work as the most efficient in the first half of the 2000s.
This is due to falling transformation costs throughout the economy but the
transaction function does not augment, which means the transaction sector is not
supportive enough and in the future its sluggishness may stifle the growth of the
economy.
Also in the second period, Taiwan suffers from the biggest drop in efficiency.
This because the transformation sector probably undergoes a price shock in its
transformation inputs. The median increase is more than 6 % points, and the major
signal from my model is that the transaction sector does not react in time, it does not
respond with a massive lowering of its output as input in other industries. So Taiwan
may not fall into the categories of Table 3 but this missing may be interpreted as
inability of the transaction and the transformation sectors to cooperate and
synchronize their developments.
The model has to be polished for several things. As for marking lines of
changes: is it the 1 % point of change that marks off a statistically significant change,
or not? Is it the 10 % points difference in the output growths that put a country in the
first block of Table 3, or in the second and the forth? Is it the median change that
determines the direction of the change, and shouldn’t we look closer at outlying
industries that may have key role in the economy?
The model cannot actually identify institutionally led structural changes, as it
came out in the course of the research, but it can identify successfully changing
countries and it incorporates matters of institutional character. It works like a
“necessary condition” not “sufficient condition” for recognizing institutional changes
that lead to more efficient structure of the economy.
7. CONCLUSIONS
In this paper I create a partial model for analyzing the results from the
interactions between institutional and technological changes that happen at
economy-wide level. It uses nominal data and this is a modification in comparison
with the original framework of Wallis and North (1994) for integration of these two
types of changes. The core idea of this quantitative model is that the evolution of the
institutions – whether they evolve efficiently or not – can be traced with transaction
costs. This type of costs is explained with a traditional microeconomics
simplification, the Robinson Crusoe economy.
The input data come from an international database and this allows reliable
inter-country comparisons. The model is tested with 20 countries for two 5-years
periods. The results reveal different patterns in the countries development, and
receive their relevant explanation according to the model. The model categories
correspond to different interactions between institutional and technological changes
in the economy. Examples of successful and unsuccessful structural changes
(Slovenia and France) are given. Cases of countries that do not fall in the categories of
the model can still be interpreted as lack of interactions between the industries that
are responsible for transmitting the technological and the institutional changes in the
economy.
The model cannot distinguish between technological and institutional
impulses for the structural changes in the economy. It can identify if the
consequences from these different impulses are synchronized and if this led to
improvement in the overall efficiency of the economy. Also, it discerns economies
with sluggish sectors that might fail the whole system in its future development.
More or less, the model works descriptively. It uses quantitative information
to cluster economic performances and to ease the institutional analysis coming
afterwards. In order to explain and to predict, this partial model should be upgraded
into the level of a Social Accounting Matrix. It needs to include information for the
types of professions/occupations in the industries and how different types of
households earn and spend their incomes. It also needs the Government sector to be
introduced.
Beside the analysis of efficiency changes, such a model could explain issues
relating to foreign trade and inflation.
References
Arrow, K. (1969) The organization of economic activity: Issues pertinent to the choice
of market versus non-market allocation. In: U.S. Joint Economic Committee
(eds.), The Analysis and Evaluation of Public Expenditure: The PPB System. U.S.
Government Printing Office, Washington, D.C., 59–73.
Chobanov, G. and H. Egbert (2007) The rise of the transaction sector in the Bulgarian
economy. Comparative Economic Studies, 49, 683-698.
Dagnino-Pastore, J.M. and P.E. Farina (1999) Transaction costs in Argentina. 3-rd
Annual Conference of the International Society for New Institutional Economics,
Washington, USA 16-18 September 1999.
Dollery, B.E. and W.H. Leong (1998) Measuring the transaction sector in the
Australian economy 1911 – 1991. Australian Economic History Review, 38 (3),
207-231.
Egbert, H., Ivanova, M. and G. Chobanov (2010) New data on the Bulgarian
transaction sector: 1997 – 2006. In: Chobanov, G., Plöhn, J. and Schellhaass, H.
(eds.), Policies of Economic and Social Development in Europe. Lang, Frankfurt,
115-126.
Egbert, H. and M. Ivanova (2009) Transaction Employment in the USA and Bulgaria
1997 to 2006. 12th International annual conference: Discrepancies in Economic and
Social Development in Europe, Sofia, Bulgaria 9-10 October 2009.
Hazledine, T. (2001) Measuring the New Zealand transaction sector, 1956-98, with an
Australian comparison. New Zealand Economic Papers, 35 (1), 77-100.
Miller, R. and P. Blair (2009) Input-Output Analysis: Foundations and Extensions (2-d
ed.). Cambridge University Press, New York
Sulejewicz, A. and P. Graca (2005) Measuring the transaction sector in the Polish
economy, 1996 to 2002. 9-th Annual Conference of the International Society for New
Institutional Economics, Barcelona, Spain 22-25 September 2005.
Wallis, J. and D. North (1986) Measuring the transaction sector in the American
economy, 1870-1970. In: Engerman S. and Gallman, R. (eds.), Long-Term Factors
in American Economic Growth. University of Chicago Press, Chicago and
London, 95-148.
Wallis, J. and D. North (1994) Integrating Institutional Change and Technical Change
in Economic History: A Transaction Cost Approach. Journal of Institutional and
Theoretical Economics, 150, 609-624.
Wang, N. (2007) Measuring transaction costs: diverging approaches, contending
practices. Division of Labor and Transaction Costs, 2 (2), 111-146.
APPENDIX 1. PERIOD MID-1990s TO EARLY 2000s
APPENDIX 2. PERIOD EARLY 2000s TO MID-2000s